The idea with Golem is that is that it aims to be a distributed supercomputer. A near top of the line supercomputer can be purchased for about $100 million. Golem is already valued at almost twice that now and I haven't seen anything to suggest that Golem will be able to keep up with an actual supercomputer any time soon, much less two. Until Golem proves that it can equal the power of a real supercomputer, I'm not willing to buy it at values that are more than the cost of a real supercomputer. Sold my Golem last week on this basis. Cool project, but don't think it deserves a higher value at the moment.